KR20030090631A - Method for improving the functions of the binary representation of mpeg-7 and other xml-based content descriptions - Google Patents

Abstract

It is an object of the present invention, in a sequence in which signal bits (flags) for the length information of an optional section and / or variable length code are defined, placed at the beginning of the generated bit stream, thereby providing data by sorting process and bit mask lookup. The way search is accelerated is, for example, to code data types that have a tree structure and are column-based. The present invention also relates to, among other things, more effectively coding the #position field by assignment of code fragments for each group of regular expressions.

Description

IMPROVING THE FUNCTIONS OF THE BINARY REPRESENTATION OF MPE-7 AND OTHER WML-BASED CONTENT DESCRIPTIONS}

The present invention relates to coding and decoding structured documents on the basis of XML, for example as provided in MPEG-7. XML (extensible markup language) is a standard for defining document structures and is used to represent structured data in text files, forming a base for, for example, the XHTML language. The documents structured on the basis of XML can be specified, for example, by a document type definition (DTD), XML-schema or multimedia description schema (DS)-also referred to as "schema" below. Based on the structured elements of the set.

For example, in [1], it is known to set up a navigation path using a binary table of a MPEG-7 file and a tree-structured code table.

The present invention relates to optimizing the coding of structured XML-documents. The primary object of the present invention is to provide a method for improving a content description based on XML, in particular a function for representing an MPEG-7-document in a binary manner, wherein the amount of data to be transmitted is increased. As little as possible, the searching process inside the coded document and the additional processing of the coded document are as simple as possible.

This object is achieved by the features of claims 1 and 4 according to the invention. The other claims relate to preferred refinements of the method according to the invention.

One conclusion obtained from the aforementioned ISO / IEC draft [1] is that the structure of an XML-document can be recorded as a data tree, in which case each description element corresponds to a node in the tree. The structure of the node is determined by the definition in the schema on which the document is based. In particular, the type and number of child elements are defined by the node structure. A prototype of such a node is shown, for example, on page 19 of [1]. The tree structure node consists of a name of an element or complex type, a field with a code word TBC (Tree Branch Code) used to refer to a child-element, and a tree branch representing a reference to the corresponding child-element. In [1] we also see that the TBC is split into two components: schema branching and location information, where the schema information is derived from elements that appear as child elements in the schema, while the location information is multiplexed. Contains information about the locations of the elements that may appear. In this case the possible types of child elements are again complex type elements which may contain child elements, or simple type elements or attributes which may not contain child elements. The length of field # position is defined by the maximum number of related elements ("maxOccurs") in the schema.

To move within a document, the TBC, SchemaBranchCodes and, in some cases, PositionCodes, are interconnected, which creates one path within the document. When the desired element is reached, the last code is inserted into the table. If the desired element cannot have additional children at all, i.e. an attribute or simple type, then the termination code is not needed and is not transmitted. In this case, attributes or simple type elements are subsequently coded and transmitted. The present invention basically consists of a column-based data type, that is, a data type derived directly or indirectly from a data type column, the possible contents of which can be specified via regular expression [4], eg For example, it can be shown in a tree structure, where the signal bits (flags) for the length information of the optional section and / or variable-length code are placed in the sequence defined according to a general method, and thus placed at the beginning of the generated bit stream. On the one hand, the processing of the information contained in the data type can be simplified by the preliminary information about the content, and the coding is done in such a way that the data retrieval can be accelerated by sorting and bit mask lookup. The invention also relates to more efficient coding of the #position field by allocating defined code fragments for an individual group of elements corresponding to the definition of the data type.

The invention is explained in detail below with reference to the embodiments shown in the drawings.

1 is a regular expression for defining a column-based data type having a number of optional sections interleaved,

FIG. 2 is a first description of the regular expression shown in FIG. 1 with an associated bit stream formed for a column-based data type,

3 is a further description of the regular expression shown in FIG. 1 with an associated bit stream for a column-based data type to illustrate the invention,

4 is a description of a regular expression for explaining the improvement of the coding effect according to the present invention.

In Figure 1, a data type based on a column with an optionally present section is described as a regular expression. Regular expressions included in the definition of a data type are specified in more detail in [4], for example. Under the regular expression are optional simple sections (1, 4 and 5) and correlated optional sections (2, 2.1, 2.1.1 or 3, 3.1, 3.1.1, 3.1.1.1, and 3.1.1.1.1). ) Is indicated by the corresponding line.

2 shows a representation FRA1 (T (S)) for a data type T (S) based on the column specified by the regular expression S of FIG. 1 in a flag numbered correspondingly to the optional section. In this case, following the corresponding query of the flag and optionally included information data, immediately following the individual flag as conventionally according to [2] so far. Side by side with the representation FRA1 (T (S)) the bit stream BS1 is indicated with each number of bits in a row manner. As shown in the foregoing embodiments, the class of authorized symbols, also referred to below as data fields, for example ＼d {2} may later be coded by a special binary representation. In this case a variable length data field, e.g. "#d +, " is coded by a variable length code, and a particular bit, e.g. every 9th bit, is for example followed by an additional 9 bits. Signal.

In existing methods for coding column-based data types, regular expressions are used to code the structure of an embodiment. However, in order to evaluate the coded data of the data type, it is possible to determine whether the section correlated with the receiver, i.e. whether a corresponding data field is included in the data stream and if so where it is included. This readout should be entered. In addition, although it is impossible for the decoder to predetermine and reserve the resources necessary for the decoding and storage of the data field, the decoder must execute this stepwise with the decoding of the bit stream.

In Fig. 3 a representation RRA2 (T (S)) for the regular expression S, modified corresponding to the invention, is shown with a corresponding additional bit stream BS2. In this embodiment, the flag SOP for the optional section is first stored according to the coding principle described in [2], but in this case at the beginning of the bit stream BS2, then the length of the variable data is variable. Information LVLC is stored.

The length information can be coded, for example, as follows: The number of consecutive bits equally one represents the number of data blocks used for data representation corresponding to the variable length. If the data block is made up of eight bits as in the above example, for example, the binary code " 1110 " specifies a data length of 24 bits, while the binary code " 10 " Specifies the length. The calculated data length is represented by " 8 * VLC_X " in this embodiment.

In the present invention, the data retrieval can be accelerated by signaling bits or flags for the optional section and / or length information of the sequence-determined and variable length code placed at the beginning of the bit stream. Because the first flag signals whether the current data can include the section to be searched for the regular expression, the entire bit stream BS2 need not be read in for this purpose. In addition, because of the information about the data field used and the length of the data field at the beginning of the bit stream, the remaining bit stream (RBS) for searching can be compared with a predetermined bit pattern.

In figure 4 the representation of the definition for the complex data type T to be coded is shown or shown alongside the individual schema branching code SBC and the individual location code #pos, as the corresponding tree branch table is constructed. . One element of a complex type element is assigned an SBC, for example in a sequence appearing in the definition, in which case the SBC is independent of in which group or how deeply it is correlated. Various codes are also assigned if elements of the same type and / or the same name appear in two or more different groups within the designation of the complex type. The schema branching code (SBC) clearly determines which elements are coded within which groups can be interconnected. Corresponding to the interconnection of the groups, a location code (#pos) assigned to the element can be created. In this case, for each group of elements to be coded, one code fragment, for example CF, is assigned, for example, in a fixed sequence starting from the outermost. In this case, one fixed bit length is used for each of the groups according to the maximum possible appearance. However, corresponding to the definition for coding the location code, a variable bit length can also be used. In the embodiment of Fig. 4, elements a to c are assigned schema branching codes in the appearance sequence. In this embodiment element a appears twice, and in the second appearance the element is assigned a different code "100" instead of "001". The field for location code (#pos) consists of the following components or code fragments:

Three bits (XXX) for determining the position inside the outer sequence-parentheses;

A multiple of 4 bits (YYYY) that determines the position within the choice-parentheses;

Multiple of four bits (ZZZZ) to determine the position under element (c).

The method according to the invention has the advantage over the method described in [3] that only locations where relevant elements can also be positioned according to the schema definition should be coded or addressed. In addition, only one location code should be passed for elements that can be described multiplely. Thereby, more effective coding of the technical structure can be achieved. A further important advantage is the advantage that the document can be extended to any respective relative depth by additional elements, which is often not possible due to the consecutively assigned location codes in the method according to [3].

The following publications are cited in this document:

[1] ISO / IEC CD 15938-1 Information technology-Multimedia content description interface: Systems, ISO / IEC JTC1 SC29 / WG11 / N3701, La Baule (France), Oktober 2000

[2] Datatype encoding improvements for BiM, Cedric Thienot, ClaudeSeyrat, Gregoire Pau, ISO / IEC TTC1 SC29 / WG11 / M6853, Pisa (Italy), January 2001

[3] BiM navigation path improvements, Cedric Thienot, Claude Seyrat, Gregoire Pau, ISO / IEC JTC1 SC29 / WG11 / M6854, Pisa (Italy), January 2001

[4] aus Internet: http://www.w3.org/TR/xmlschema-2/#regexs

Claims (5)

As a method for improving the function for representing XML-based content descriptions in a binary manner, The signaling bits (SOPs) for optional sections and / or length information (LVLC) of codes with variable lengths are set in order to describe column-based data types in a fixed sequence derived from a regular expression. How a column-based data type (T (S)), defined by a regular expression (RRA2 (T (S))), is tabulated. The method of claim 1, First, the data of the bit stream in the at least one description is retrieved in such a way that the signaling bit includes a specific optional section of the regular expression. The method of claim 1, A method in which the data of a bit stream in at least one description is retrieved in such a way that signaling bits are first used for the purpose of preparing a corresponding resource in the decoder in advance for processing and / or storing the contained data field. The method of claim 2 or 3, Using the length information, a bit pattern for searching in the remaining bit stream (RBS) is formed. As a method for improving the function for representing XML-based content descriptions in a binary manner, Various tree branching codes (SBCs) are assigned to structural elements in various description groups. Location code (#pos) for the various connection depths at which the structural element occurs comprises different code fragments (CF).